Efficient Computation of Equilibria for Extensive Two-Person Games

TL;DR: In this article, the formation of paths in an information network is modeled as the Nash equilibrium of an N player routing game and the task of obtaining this equilibrium is shown to be NP-Hard.

TL;DR: Theoretical results for CFR are extended when using function approximation, and worst-case guarantees with function approximation are complemented with experiments on several common benchmark games with sequential decision making and imperfect information.

TL;DR: This contribution proposes a framework based on a sparse zero-sum game which performs a stable functional feature selection based on feature subsets ranking by a thresholding stochastic bandit and provides a theoretical analysis of the introduced algorithm.

TL;DR: In this article, a game-theoretic centralized training regimen is proposed to effectively perform trajectory sampling so as to foster team coordination in a zero-sum, imperfect-information game, where each team member's policy is parametrized as a neural network whose output is conditioned on a suitable exogenous signal.

TL;DR: An algebraic proof of the existence of equilibrium points for two-person non-zero-sum games is given in this paper, leading to an efficient scheme for computing an equilibrium point, which is valid for any ordered field.

TL;DR: In this paper, simple constructive proofs are given of solutions to the matric matric system Mz − ω = q; z ≧ 0; ω ≧ 1; zT = 0, for various kinds of data M, q, which embrace quadratic programming and the problem of finding equilibrium points of bimatrix games.

TL;DR: Several new complexity classes of search problems, ''between'' the classes FP and FNP, are defined, based on lemmata such as ''every graph has an even number of odd-degree nodes.''